
A funder report arrives with a column you have to fill in called "explanation of variance," and the honest answer is often that nobody looked until the report was due. By then the difference between a timing issue and a real overspend is hard to reconstruct, and the explanation you write sounds like a guess because it is one. A grant variance report is not your organisational budget report with one grant pulled out of it. It answers to an agreement, tracks against approved budget lines, and runs on the funder's period. This guide covers how to build it, how to read it, and how to explain it.
It reports against a contract, not against a plan. Your organisational budget is a decision the board made and can revise. A grant budget is a term of an agreement, and moving money between its lines may need permission.
That difference changes three things. The comparison period is the funder's, not yours. The line items are the funder's, which rarely match your account structure exactly. And an unexplained variance is a compliance matter rather than a management one.
It also changes who reads it. A board reads variance as information. A funder reads it as evidence about whether your organisation does what it said it would do with money.
Treat the two reports as separate products built from the same ledger. Nothing about a grant variance report should require a second set of books.
From the grant dimension, filtered to the funder's period and mapped to the funder's budget lines. If you are exporting to a spreadsheet and typing numbers in, the report is no longer tied to your books.
Three things have to be in place. Every grant-funded transaction carries the grant code at the time it is entered. The mapping from your account codes to the funder's budget lines is written down and reused. And the report is run from the accounting system rather than assembled beside it.
Keep the mapping document with the grant file. It is the thing that lets a different person produce next quarter's report and get the same answer.
Where a funder's budget line has no natural home in your chart of accounts, resist creating one. Map two or three of your accounts to it instead and record the mapping. Accounts should serve the whole organisation, not one funder.
Timing variances close on their own; permanent ones do not. Telling them apart is most of the analysis, and it changes what you write.
Work through that order every time. The fourth row is the one to check first, because a classification error dressed up as a timing variance is the version that survives into a final report and gets found by someone else.
A useful test: ask what the line will look like at the end of the grant period if nothing changes. If it converges, it is timing. If it does not, it is permanent and needs a decision rather than a note.
A cost incurred inside the period but invoiced after it still belongs in the period for your ledger, so a report built only on bank activity can show an underspend that is not real. Whether that cost can go into the claim is a separate question the agreement and the claim form answer.
Three checks catch most of it. Look for goods and services received before the period end where the invoice arrived later. Check payroll for a pay period that straddles the cutoff. And look for anything paid in the period that relates to the next one, such as an annual subscription or a prepaid deposit, which pushes the variance the other way.
Post accruals in your ledger for the ones that belong, with enough description that the entry explains itself in six months. Whether those accrued costs can go into the claim is a separate question, and the agreement and the claim form decide it. Some funders accept incurred but unpaid costs on progress claims while requiring a final claim to cover only amounts actually paid. Check before submitting rather than assuming your accounting basis carries over.
This is also where a monthly review earns its cost. Accruals reconstructed at the end of a two-year grant are estimates. Accruals posted monthly are records.
Read the reallocation clause before you need it, because most set a threshold and a notice period. It is one of the most useful clauses in a grant agreement and one of the least read.
Agreements commonly allow movement between budget lines up to a percentage of the line or the total without approval, and require written approval above it. Some require notice a set number of days in advance. Some prohibit any movement into or out of named categories, salaries and capital being the usual ones.
Copy the clause and its threshold into your grant register when the agreement is signed. Then the monthly review compares the running variance against the threshold rather than against an instinct about what the funder will accept.
Where you need approval, ask early and in writing. A request made in month four with a plan attached reads differently from one made in month eleven with a completed overspend attached.
Name the cause, quantify it, and say what happens next. A single sentence that does all three beats a paragraph that does none.
A workable explanation has four parts: what the variance is in dollars and as a percentage of the budget line, why it happened in concrete terms, whether it is timing or permanent, and what you are doing about it. "Program supplies are $4,200 under budget, or 28%, because the spring session was deferred to the autumn term. We expect the line to be fully spent by the end of the grant period."
Avoid three habits that make reviewers ask follow-up questions. Do not explain a variance by restating the number in words. Do not attribute a variance to "timing" without saying what was delayed. And do not report a favourable variance as a success when it means the funded activity did not happen.
Where the variance reflects a change in how the program ran, say so plainly. Funders generally deal much better with a described change than with a discovered one.
Keep the explanations you write. Repeated variances on the same line across periods signal that the budget was wrong rather than the spending, which is a useful thing to know when the next agreement is negotiated.
Update the budget in your system and keep the approval with the grant file, so the report and the agreement never disagree. A reallocation that lives only in an email thread will be missed.
When approval comes through, revise the approved budget by line in your grant register with the date and the reference to the approval. Your comparison from that point runs against the revised figures. Retain the original budget too, so the history is visible.
Do not silently overwrite prior reports. The reports issued during the period reflected the budget in force at the time, and a funder reviewing the file wants to see the sequence rather than a tidy version of it. Retain the originals, and amend or reissue a report only where the funder or the agreement requires or permits it, since some funders do provide for corrected claims.
Where a funder declines a reallocation, the overspend has to be funded from unrestricted resources and coded out of the grant. Doing that as a dated journal entry with a reason is the difference between a clean file and an awkward conversation.
Grant-level reporting is much easier when the underlying structure was built for it. Our guide to fund accounting covers how restricted balances stay visible in the ledger, and our list of common pitfalls covers what else tends to break. Where reporting is consuming more time than the program, ongoing bookkeeping support is usually cheaper than the reconstruction.
Monthly, even where the funder only asks quarterly or annually. The point is not the reporting deadline; it is having time to act on what the variance shows. A monthly review also means the explanation is written while someone still remembers the cause.
Your agreement usually sets the threshold that matters, and it should be recorded in your grant register when the agreement is signed. Internally, many organisations review anything beyond ten percent of a budget line, but the contractual threshold is the one with consequences.
That depends on the agreement, so read it first. Options generally include reallocating within the budget where permitted, requesting an extension of the grant period, or returning the unspent balance. Raising it early gives you all three; raising it at the end usually leaves the last one.
Yes, and they often matter more. An underspend usually means planned activity did not happen, which is a program question as much as a financial one. Funders read an unexplained underspend as a delivery problem.
Map rather than restructure. Record which of your accounts roll up to each funder budget line, keep the mapping with the grant file, and reuse it. Rebuilding your chart of accounts around one funder creates a bigger problem than it solves.
Whoever is accountable for delivering the program, working with whoever produces the numbers. A bookkeeper can say the line is underspent by $4,200; only the program lead can say why. Getting both names on the review is what makes the explanation reliable.
This article covers grant-level variance tracking and reporting as at September 2026. Reallocation rights, thresholds and reporting requirements are set by each grant agreement and vary considerably, so read yours before acting on anything described here. It is not advice on any particular agreement.